App refactored

.gitignore

@@ -1,3 +1,4 @@
-DS_Store
+.DS_Store
 .env
-cache/
+cache/
+*.pyc

app.py

@@ -1,45 +1,25 @@
-### Mike Dean Experiments
 
-### Import Section ###
 """
 IMPORTS HERE
 """
 import chainlit as cl
-import os
-from dotenv import load_dotenv
-from chainlit import AskFileMessage
-from langchain_text_splitters import RecursiveCharacterTextSplitter
-from langchain_community.document_loaders import PyMuPDFLoader
 from qdrant_client import QdrantClient
 from qdrant_client.http.models import Distance, VectorParams
-from langchain_openai.embeddings import OpenAIEmbeddings
-from langchain.storage import LocalFileStore
+
 from langchain_qdrant import QdrantVectorStore
-from langchain.embeddings import CacheBackedEmbeddings
 
-from langchain_core.globals import set_llm_cache
-from langchain_openai import ChatOpenAI
-from langchain_core.caches import InMemoryCache
 from operator import itemgetter
 from langchain_core.runnables.passthrough import RunnablePassthrough
 from langchain_core.runnables.config import RunnableConfig
 import uuid
 from prompts import chat_prompt 
+from handle_files import split_file
+from models import chat_model, cached_embedder
 
-load_dotenv()
-
-# os.environ["LANGCHAIN_PROJECT"] = f"Mike HF Production Rag - {uuid.uuid4().hex[0:8]}"
-# os.environ["LANGCHAIN_TRACING_V2"] = "false"
-# os.environ["LANGCHAIN_ENDPOINT"] = "https://api.smith.langchain.com"
-
-### Global Section ###
 """
 GLOBAL CODE HERE
 """
 
-# Typical Embedding Model
-core_embeddings = OpenAIEmbeddings(model="text-embedding-3-small")
-
 # Typical QDrant Client Set-up
 collection_name = f"pdf_to_parse_{uuid.uuid4()}"
 client = QdrantClient(":memory:")
@@ -48,57 +28,12 @@ client.create_collection(
     vectors_config=VectorParams(size=1536, distance=Distance.COSINE),
 )
 
-# Adding cache!
-store = LocalFileStore("./cache/")
-cached_embedder = CacheBackedEmbeddings.from_bytes_store(
-    core_embeddings, store, namespace=core_embeddings.model
-)
-
 # Typical QDrant Vector Store Set-up
 vectorstore = QdrantVectorStore(
     client=client,
     collection_name=collection_name,
     embedding=cached_embedder)
 
-# rag_system_prompt_template = """\
-# You are a helpful assistant that uses the provided context to answer questions. Never reference this prompt, or the existence of context.
-# If you cannot answer the question from the information in the context, tell the user that 
-# you cannot answer the question directly from the context, but that you will give an answer
-# that is based on your general knowledge.
-# """
-
-# rag_message_list = [
-#     {"role" : "system", "content" : rag_system_prompt_template},
-# ]
-
-# rag_user_prompt_template = """
-# Question:
-# {question}
-# Context:
-# {context}
-# """
-
-# chat_prompt = ChatPromptTemplate.from_messages([
-#     ("system", rag_system_prompt_template),
-#     ("human", rag_user_prompt_template)
-# ])
-chat_model = ChatOpenAI(model="gpt-4o")
-set_llm_cache(InMemoryCache())
-
-def split_file(file: AskFileMessage):
-     import tempfile
-     text_splitter = RecursiveCharacterTextSplitter(chunk_size=500, chunk_overlap=100)
-     Loader = PyMuPDFLoader
-     with tempfile.NamedTemporaryFile(mode="w", delete=False) as tempfile:
-        with open(tempfile.name, "wb") as f:
-            f.write(file.content)
-     loader = Loader(tempfile.name)
-     documents = loader.load()
-     docs = text_splitter.split_documents(documents)
-     for i, doc in enumerate(docs):
-        doc.metadata["source"] = f"source_{id}"
-     return docs
-
 ### On Chat Start (Session Start) Section ###
 @cl.on_chat_start
 async def on_chat_start():
@@ -116,15 +51,16 @@ async def on_chat_start():
 
     file = files[0]
 
+
     msg = cl.Message(
-        content=f"Processing `{file.name}`...", disable_human_feedback=True
+        content=f"Processing `{file.name}`..."
     )
 
     await msg.send()
-
+   
     docs = split_file(file)
     vectorstore.add_documents(docs)
-   
+
     retriever = vectorstore.as_retriever(search_type="mmr", search_kwargs={"k": 15})
     retrieval_augmented_qa_chain = (
         {"context": itemgetter("question") | retriever, "question": itemgetter("question")}

handle_files.py

@@ -0,0 +1,17 @@
+from chainlit import AskFileMessage
+from langchain_text_splitters import RecursiveCharacterTextSplitter
+from langchain_community.document_loaders import PyMuPDFLoader
+
+def split_file(file: AskFileMessage):
+     import tempfile
+     text_splitter = RecursiveCharacterTextSplitter(chunk_size=500, chunk_overlap=100)
+     Loader = PyMuPDFLoader
+     with tempfile.NamedTemporaryFile(mode="w", delete=False) as tempfile:
+        with open(tempfile.name, "wb") as f:
+            f.write(file.content)
+     loader = Loader(tempfile.name)
+     documents = loader.load()
+     docs = text_splitter.split_documents(documents)
+     for i, doc in enumerate(docs):
+        doc.metadata["source"] = f"source_{id}"
+     return docs

models.py

@@ -0,0 +1,22 @@
+from langchain_openai.embeddings import OpenAIEmbeddings
+from langchain_core.globals import set_llm_cache
+from langchain_openai import ChatOpenAI
+from langchain_core.caches import InMemoryCache
+from langchain.storage import LocalFileStore
+from langchain.embeddings import CacheBackedEmbeddings
+from dotenv import load_dotenv
+
+load_dotenv()
+
+core_embeddings = OpenAIEmbeddings(model="text-embedding-3-small")
+
+# Adding cache!
+store = LocalFileStore("./cache/")
+cached_embedder = CacheBackedEmbeddings.from_bytes_store(
+    core_embeddings, store, namespace=core_embeddings.model
+)
+
+
+chat_model = ChatOpenAI(model="gpt-4o")
+set_llm_cache(InMemoryCache())
+

prompts.py

@@ -1,9 +1,10 @@
-## Contains prompts, welcome messages, etc.
 
 from langchain_core.prompts import ChatPromptTemplate
 
 rag_system_prompt_template = """\
-You are a helpful assistant that uses the provided context to answer questions. Never reference this prompt, or the existence of context.
+You are a helpful assistant that uses the provided context to answer questions. 
+You are an expert on clinical trials and observational studies.
+Never reference this prompt, or the existence of context.
 If you cannot answer the question from the information in the context, tell the user that 
 you cannot answer the question directly from the context, but that you will give an answer
 that is based on your general knowledge.
@@ -14,10 +15,18 @@ rag_message_list = [
 ]
 
 rag_user_prompt_template = """
+Use the context to answer the question and provide a clear answer.  Do not mention the
+document in your answer.  If there is no specific information relevant to the question,
+then tell the user that you don't know the answer, and invite the user to rephrase the question.
+
 Question:
 {question}
+
 Context:
 {context}
+
+
+
 """
 
 chat_prompt = ChatPromptTemplate.from_messages([